I like to think of an agent as a LLM chat loop with tool calling. The agentic behavior emerges with these ingredients.

Don’t really a harness…unless you consider a shell script that runs that loop a harness.

I’m also suspicious. Seems like OpenAI wants a Mythos moment too.

131 billion for me

Congrats on launching!

I find these to be astute and fascinating observations. Thanks for the read!

Maybe next step is have the llm create persistent tools from the queries it uses most often.

I thought about doing that, or having the LLM create and save HTML components, but for this particular experiment I wanted to keep it as pure and unfiltered as possible.

I've gone down this line of thought, but after adding cache lines that are highly problematic, I just end up back to LLM generating regular code as normal development calls for.

What is their success rate? What is "pretty high?"

We don't know. It's classified. That's the part of the point. I get why the OC might assume the success rate is pretty high because the successful missions can be talked about so the public hears about them. Very rare to hear about the failures like this one. But we just don't know if its rare because they don't happen much or because they happen a lot and are kept as a state secret.

I’m interested in Solace but notice I don’t see any salary ranges posted. Please consider attaching these to the job postings.

I have tinnitus from going to loud concerts. I want to evangelize the importance of wearing ear plugs. It’s just not worth having hearing damage.

Yep. Had a DJ mate who religiously wore plugs to every night, wise dude.

Wear ear plugs at gigs, kids. Older you will thank you.

There are animal markets all over China. This isn’t just any market. It’s a market that happens to be near a virology lab that studies bat coronaviruses, which had been investigated for questionable safety practices by u.s. inspectors.